Neural Networks with TensorFlow in Python

Group's Website

As a summer activity in our lab, we have prepared a small internal workshop to share experiences and speed up our learning with Tensorflow which is an open source software for machine learning capable of building and training neural networks to detect and decipher patterns and correlations.

• Date: Friday, May 26, 2017 from 9:30 am to 12:30 pm.
• Location: Phillips 149 (Mason Seminar Room), 361 Physics Rd, East Lansing, MI 48825
• Material needed: Bring your laptop. The software will be installed in your own laptop and it will be run either in the HPCC or in your laptop.

Schedule

1. 9:30-9:50. Brief introduction to Neural Networks & DNN in the context of our research activities

• Slides
• Handout

2. TensorFlow

   • 9:50-10:40. Introduction to the software.
     • Presentation (video) (4:03)
     • General aspects (video) (5:37)
     • Basic elements (video) (4:50)
     • Simple Linear Model (video) (4:23)
   • 10:40-11:00. Installation in a PC (by Marco). (Handout)
   • 11:00-11:20. Tensorflow in the HPCC (by Mengying). (Handout)

3. Examples with TensorFlow

   • Hypothetical Neural Network example.
   • 11:20-11:35. Genomic prediction in plants (by Christina).
   • 11:35-11:50. Prediction of phenotypes using image data (by Marco).
     • Slides
     • Python code 1
     • Python code 2
   • 11:50-12:05. Genomic prediction in humans (by Mengyin). (Slides)

4. 12:30. Lunch in the Gallery at Snyder Phillips.

---